Figure 2 | Scientific Reports

Figure 2

From: Deep learning to detect macular atrophy in wet age-related macular degeneration using optical coherence tomography

Figure 2

Performance on different sizes of training data. There are 2211 images in total, with 80% of images (1784) for training, 10% of images (221) reserved for validation, and the rest 10% of images (316) for testing. Among these, 1784 images were separated randomly based on different percentages for training. A plateau was reached at 0.706 with an 80% sample size.

Back to article page